VMware Cloud Foundation Installation — Deployment Wizard & JSON Spec
Objectives
- Access and configure VCF Installer appliance (online/offline depot)
- Explain VCF Fleet concept (Operations + Automation + Instances)
- Walk through the 11-step deployment wizard configuration sequence
- Select appropriate DVS profile based on NIC count (2/4/6)
- Use JSON deployment specification for repeatable deployments
- Describe the 8-step installation workflow sequence
- Identify key log files for troubleshooting VCF installation
Prerequisites
Active Holodeck VCF 9.0 lab or access to VCF documentation
Required skills:
- Basic VMware terminology
Tasks
Task 1 VCF Installation — Wizard & JSON Methods
Master both VCF installation methods with detailed configuration sequence
VCF Installer = the VCF-SDDC-Manager-Appliance OVA in installer mode (p.1265). Two documented placements: outside the management infrastructure (reusable to deploy several platforms), or on one of the future management-domain hosts — in that case it is Storage-vMotioned to the new datastore and switches into SDDC Manager mode, so it cannot be reused. Access via FQDN with admin@local credentials. Pre-deployment: (1) Review prerequisites, (2) Configure depot settings — online (Broadcom credentials) or offline (Binary Transfer Utility). Binaries NOT included in appliance — must be downloaded.
VCF Fleet = VCF Operations + VCF Automation + 1+ VCF Instances. Each VCF Instance must have a management domain. Can reuse existing VCF Operations instance. Can convert existing vCenter into management domain vCenter. Deployment models: Simple (1 NSX node, 1 VCF Ops, 1 VCF Auto — NOT production) vs HA (3 nodes each — VIP auto for NSX, no LB for VCF Ops from installer, VCF Auto handles LB internally).
Step-by-step: (1) General info — instance name, mgmt domain name, deployment model, DNS, NTP; (2) VCF Operations — appliance details, fleet mgmt appliance, collector appliance; (3) VCF Automation — FQDN, primary IP, secondary IP, CIDR for container communication (4 IPs for 3-node HA); (4) vCenter — FQDN, datacenter name, cluster name, domain, appliance size; (5) NSX Manager — appliance size, FQDN, admin/root/audit passwords; (6) Storage — vSAN (OSA/ESA, FTT, dedup/compression) or NFS or VMFS-FC; (7) Hosts — up to 64 ESX hosts, fingerprint verification; (8) Networks — VLAN, CIDR, MTU, gateway for mgmt/VM mgmt/vMotion/vSAN; (9) DVS — profiles based on NIC count (2/4/6 NICs); (10) SDDC Manager — appliance config; (11) Review → Validate → Deploy.
Profiles based on physical NIC count: 2 NICs → default/unified fabric (single VDS, all traffic), 4 NICs → Storage Traffic Separation (2 VDS, vSAN on separate NICs) OR NSX Traffic Separation (2 VDS, NSX on separate NICs), 6 NICs → Storage + NSX Traffic Separation (3 VDS — mgmt/vMotion/VM on VDS1, vSAN on VDS2, NSX on VDS3).
Alternative to wizard: upload JSON spec file containing all config parameters. JSON can be generated from wizard inputs (download after completing wizard) or created manually. After upload: preview in JSON format, edit via wizard, validate, deploy. Useful for repeatable deployments and automation.
Holodeck automates the same workflow that VCF Installer performs manually. For VCF 9.x, New-HoloDeckInstance deploys the VCF Installer appliance and drives it with a JSON deployment spec (Cloud Builder is used only for VCF 5.2.x targets). The config.json maps directly to the deployment specification JSON — same parameters, same validation. Understanding both methods reinforces VCF installation concepts.
[HOLODECK NOTE] Holodeck deployment via New-HoloDeckInstance automates the VCF Installer + JSON spec process. Key difference: Holodeck pre-configures networking (VLANs, port groups) on the physical host's vSwitch, so DVS profile selection in the wizard maps to virtual NICs, not physical pNICs. The 2/4/6 NIC DVS profiles in production correspond to physical NIC separation for management, vSAN, vMotion, and NSX traffic — in Holodeck all traffic shares the physical host's NICs.
Validation Gate
Check: List the five most critical pre-validation checks the VCF Installer performs before starting VCF bring-up
Expected: 1. DNS forward/reverse resolution for all hostnames. 2. NTP reachability and time sync. 3. ESX host HTTPS reachability, with each host certificate common name matching the FQDN given in the deployment wizard. 4. Network connectivity between hosts. 5. Password complexity validation.
Common Errors
Task 2 VCF Installation Workflow & Troubleshooting
Understand the sequential installation workflow and key log files
Ordered sequence: (1) ESX Hosts Preparation — interactive install, mgmt VLAN, VM VLAN, NTP config; (2) vCenter Deployment; (3) SDDC Manager Deployment; (4) vSphere Cluster Configuration — hosts migrated from VSS to VDS; (5) NSX Deployment and Configuration; (6) VCF Operations Fleet Management Deployment; (7) VCF Operations Deployment and Configuration; (8) VCF Automation Deployment and Configuration.
Installation progress visible in VCF Installer UI Tasks section. Detailed log files for each phase of installation. For troubleshooting: identify which workflow step failed, examine corresponding logs, check for config validation errors. Documented locations (VCF 9.0 docs, troubleshooting the VCF Installer): validation or deployment failures — /var/log/vmware/vcf/domainmanager/domainmanager.log on the VCF Installer appliance; VCF Operations or VCF Automation deployment issues — /var/log/vrlcm/vmware_vrlcm.log on the VCF Operations fleet management appliance.
Validation Gate
Check: What is the correct order of component deployment during VCF bring-up?
Expected: ESX host preparation and validation → vCenter deployment → SDDC Manager → cluster/VDS configuration → NSX Manager deployment and configuration → VCF Operations fleet management → VCF Operations → VCF Automation (matches Task 2 step 1; confirm against the VCF Installer task list of a completed deployment — no published order surfaced in the local docs)
Common Errors
Design Reflection (VCDX)
VCF bring-up is a Day-0 operation that VCDX candidates should understand end-to-end. Panelists may ask about failure scenarios during bring-up and your recovery strategy. Know the deployment sequence and where common failures occur.
Requirements
- Understand the VCF Installer deployment wizard and JSON specification
- Know the bring-up component deployment sequence
- Identify pre-validation checks and common failure points
Constraints
- All DNS records must exist before bring-up
- All hosts must be reachable over HTTPS on the management network, with certificates whose common name matches the host FQDN
- Password complexity must meet VCF requirements
Assumptions
- Network infrastructure (switches, VLANs, routing) is pre-configured
- DNS and NTP services are operational and reachable
Risks
- IP conflicts in JSON spec causing mid-deployment failure
- DNS misconfiguration blocking pre-validation
- Partial deployment failure requiring manual cleanup